Accelerat ing t he world's research.

A mathematical formulation for a

school bus routing problem
Patrick Schittekat

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

A mat heurist ic for t he school bus rout ing problem

Marc Sevaux

T he urban bus rout ing problem in t he Tunisian case by t he hybrid art ificial ant colony algorit hm
Rafaa Mraihi

A real geographical applicat ion for t he School Bus Rout ing Problem
Crist iano M Silva
 A mathematical formulation for a school bus routing problem

Patrick Schittekat1 , Marc Sevaux2 , Kenneth Sörensen3∗

1
University of Antwerp, Belgium (patrick.schittekat@ua.ac.be)
2
University of South Brittany, France (marc.sevaux@univ-ubs.fr)
3
University of Antwerp, Belgium (kenneth.sorensen@ua.ac.be)

ABSTRACT

The school bus routing problem discussed in this paper, is similar to the standard vehicle routing problem, but has several
interesting additional features. In the standard VRP all stops to visit are given. In our school bus routing problem, we assume
that a set of potential stops is given, as well as a set of students that can walk to one or more of these potential stops. The
school buses used to pick up the students and transport them to their schools have a finite capacity. The goal of this routing
problem is to select a subset of stops that will actually be visited by the buses, determine which stop each student should
walk to and develop a set of tours that minimize the total distance travelled by all buses. We develop an integer programming
formulation for this problem, as well as a problem instance generator. We then show how the problem can be solved using a
commercial integer programming solver and discuss some of our results on small instances.
Keywords: Schoolbus, routing, stop selection, mathematical programming, exact algorithm

1. INTRODUCTION

This research is motivated by a real-life school bus rout-

ing problem. In the Flemish region, transport of children
to their school is organized by the Flemish transportation
society. Each pupil that has the right to be transported to its
school, also has the right to a bus stop located at a distance
of less than 750 metres from its residence. To efficiently
design the routes, a set of potential stops is determined, so Fig. 1: Unsolved problem
that each pupil has at least one stop he/she can walk to.
Routes are then determined for the school buses so that all
students are picked up at a stop they are allowed to use,
while making sure that the capacity of the buses is not ex-
ceeded.

Contrary to most vehicle routing formulations, in which a

set of stops is given and routes need to be determined that
visit each stop, we discuss in this paper a vehicle routing
formulation in which a set of potential stops is given, but in
which determining the set of stops to actually visit is a part Fig. 2: A possible solution
of the problem formulation. The objective of this problem
is to simultaneously (1) find the set of stops to visit, (2)
determine for each customer which stop he should move to tomer has a certain demand and the vehicles have a certain
and (3) determine routes that visit the chosen stops, so that capacity, that cannot be exceeded.
the total travelled distance is minimized.
We assume that the students cannot walk directly to the
Figure 1 shows an example of such a school bus routing school. Students that can walk to school do not have to be
problem. In this figure, dots represent students, the small taken into account and may consequently be removed from
squares represent potential stops, the large square repre- the problem data.
sents the school. Dotted lines indicate which stop a student
is allowed to walk to. The assignment of students to stops used by the routes is by
itself a difficult problem. When students can be assigned to
Assuming that the capacity of the buses is 8, a possible (but more than one stop in the same route, the assignment to a
not necessarily optimal) solution to this problem is shown stop is arbitrary. This is not necessarily the case if a student
in figure 2. can be assigned to more than one stop in different routes.
In this case, students should be assigned to stops in such
In the problem discussed in this paper, we assume that all
a way that the capacity of the vehicles is not violated. In
students represent a unit to be transported and that the ca-
figures 1 and 2, there is one student that can move to a stop
pacity of the buses can be expressed as a whole number of
in both of the routes. However, given that the capacity of
units. In a more general version of this problem, closer to
the buses is 8, this student needs to be assigned to the route
the classical capacitated vehicle routing problem, each cus-
on the right.
∗ 1-4244-0451-7/06/$20.00 2006
c IEEE. Next to the obvious school bus routing application, this

1552
problem formulation has other applications. For example, velop a multi-objective optimization problem, one of the
some parcel delivery services nowadays offer the option of objectives being the minimization of the total walking dis-
delivering at a set of pre-defined drop-off points. This has tance by all students.
obvious cost-saving advantages over delivering at any lo-
cation specified by the customer. Customers have to de- 3. PROBLEM FORMULATION
cide beforehand at which drop-off point they wish to pick
up their items. It can be envisaged that customers may be As should be clear from the previous discussion, most pa-
asked to specify more than one drop-off point and that the pers on the routing of school buses have focussed on cap-
parcel delivery company will then choose among the ones turing as many aspects of reality as possible, building in-
chosen by at least one customer in such a way that rout- tricate multi-objective models with many complex con-
ing costs are minimized but every customer can pick up straints. In this paper, we focus on a very basic version
his parcel at one of the drop-off points he specified. Cus- of this problem to develop efficient methods that take the
tomers may e.g. be notified by a mobile phone message structure of the problem into account. We focus specifi-
of the drop-off points their package will be delivered. In a cally on the stop selection aspect. Our problem formula-
more complex setting, the price of the delivery may depend tion has only a single school, one type of student and one
on the number of drop-off points specified by the customer. type of bus, with fixed capacity. We optimize the standard
vehicle routing criterion: the total distance travelled by all
The objective of this paper is to define the school bus rout- vehicles.
ing problem as a mixed integer linear programming prob-
lem and solve it using both exact and heuristics methods. The schoolbus routing problem can be expressed as an inte-
As will become clear from the literature review, previous ger linear programming problem. We assume that the graph
research has focused mainly on building intricate multi- on which the problem is defined, is directed. The follow-
objective models of similar problems. The focus on this ing formulation builds upon the one discussed on page 15
paper is on understanding the problem in its simple form of Toth and Vigo [9].
and building a powerful meta-heuristic to solve large in-
3.1. Data
stances quickly.
cij Cost of traversing arc ij
2. LITERATURE REVIEW K Number of buses
C Capacity of the buses
Contrary to the literature on the ordinary vehicle routing V Set of all potential stops
problem and several of its extensions (e.g. time windows), E Set of all arcs between stops
only a limited amount of research has considered the rout- S Set of all students
ing of school buses. In the literature on routing of school sli Binary variable that indicates whether student l
buses, only a small number of papers considers the simul- can walk to stop i or not
taneous selection of stops.
3.2. Decision variables
Most school bus vehicle routing formulations focus on for-
mulating extra constraints and/or objectives to take some xijk number of times vehicle k traverses arcs
student-related factors into account. Bodin and Berman [2], from i to j
Braca et al. [4], and Desrosiers et al. [6], add a maximum yik 1 if vehicle k visits stop i, 0 otherwise
travel-time constraint for each student and/or a time win- zilk 1 if student l is picked up by vehicle k at stop i,
dow for arrival at the school. Bennett and Gazis [1] add the 0 otherwise
total travel time of all children as an objective.

Thangiah et al. [8] discuss the routing of school buses in ru-

ral areas. They develop a system that is able to solve large-
scale routing problems with a large number of complex XX X
K

constraints and several objectives. Interestingly, the au- min cij xijk (1)
thors note that local government subsidizing policies may i∈V j∈V k=1
result in very ineffective routings, e.g. maximizing the time X
K
that students spend on a bus instead of minimizing this. The y0k ≤ K, k = 1, . . . , K (2)
authors do not consider automatic stop selection in their k=1
problem formulation. X X
xijk = xjik = yik , ∀i ∈ V, k = 1, . . . , K (3)
Only a limited number of papers on school bus routing con- j∈V j∈V
XX
siders the automatic selection of bus stops to be a part of the xijk ≥ yhk ,
optimization problem. In Dulac et al. [7], students are as-
i∈S j ∈S
/
signed to an intersection of streets adjacent to the street of
their residence. A subset of these potential bus stops is then ∀S ⊆ V \ {0}, h ∈ S, k = 1, . . . , K (4)
selected and a VRP solved. In Chapleau et al. [5], potential X
K
stops are first clustered, after which stops are selected so yik ≤ 1, ∀i ∈ V \ {0} (5)
that a maximum number of students has a stop within walk- k=1
ing distance. The school bus routing problem discussed in
X
K
Bowerman et al. [3] includes a maximum walking distance zilk ≤ sli , ∀l ∈ S, ∀i ∈ V (6)
for a student to his/her assigned bus stop. The authors de- k=1

1553
 XX
zilk ≤ C, ∀k = 1, . . . , K (7)
i∈V l∈S
zilk ≤ yik , ∀i, l, k (8)
XX
K by first generating for each student j an angle αj ∈ [0, 2π]
zilk = 1, ∀l ∈ S (9)
i∈V k=1
yik ∈ {0, 1}, ∀i ∈ V, k = 1, . . . , K (10)
xijk ∈ {0, 1}, ∀i, j ∈ V |i 6= j (11)
zilk ∈ {0, 1}, ∀i, j ∈ V |i 6= j (12)
The objective function (1) minimizes the total travelled dis-
tance by all vehicles. Constraints (2) ensure that all ve-
hicles start from the depot. Constraints (3) ensure that if
stop i is visited by vehicle k, then one arc should be tra-
versed by vehicle k entering stop i and leaving stop i. Ca-
pacity cut constraints (4) ensure that each cut (V \ S, S)
defined by a customer set S is crossed by a number of
arcs not smaller than r(S), the minimum number of ve-
hicles needed to serve set S. These constraints serve as
subtour elimination constraints. The following constraints
are added or adjusted in comparison with the original for-
mulation to cope with the special features of this specific
vehicle routing problem. Constraints (5) ensure that all are
visited no more than once, except the stop corresponding
to the depot. Constraints (6) ensure that each student walks
to a single stop he or she is allowed to walk to. Constraints
(7) ensure that the capacity of the vehicles is not exceeded.
Constraints (8) ensure that student l is not picked up at stop
i by vehicle k if vehicle k does not visit stop i. Constraints
(9) ensure that all students are picked up once. Finally,
constraints (10), (11), and (12) ensure that all decision vari-
ables are binary. This corresponds to ensuring respectively
that a vehicle k either visits a stop i or it does not, a vehicle
k either drives from one stop i to another stop j or it does
not, and a vehicle k either picks up a student l at stops i or
it does not.
(xmax , ymax ).
For each generated stop, ns student positions are generated
at a distance of maximum wmax from the stop. This is done
and a distance wj from the stop. The student is then put at
(x, y)-coordinates equal to (xi +wj cos αj , yj +wj sin αj ).
5. BENCHMARK EXACT SOLUTIONS
An exact algorithm using the IP formulation proposed in
this paper has been implemented in a commercial IP solver
and is currently used to solve (small) benchmark instances.
Preliminary results are reported using a Pentium 4, 3.20
GHz running linux. All the source code is written using the
commercial ILP modelling language Xpress-Mosel 1.6.0
and solved with the optimizer 16.10.07 from Dash Asso-
ciates. Only three instances have been generated for this
first experimental part.
As usual, the IP model is first solved ignoring the breaking
subtour constraints (4). Once an optimal solution of the
subproblem is obtained, i.e. a lower bound of the global
problem, breaking subtour constraints are added one by
one and the subproblem with these additional constraints
is solved again. This procedure is repeated until no illegal
subtours are found. The solution obtained after all these
iterations is one of the optimal solution of the problem.
A typical execution is shown on an example with 10 stops
and 50 students given on Figure 3. As in Figure 1, the depot
is represented with a large square (centered in the figure),
stops are represented with small squares and circles locate
the students. Dotted lines indicates to which bus stop a
student can walk.

By using this formulation, we implicitly make a number of

assumptions. One assumption is that a stop is only visited
by one bus. This means that the number of students per stop
may not exceed the capacity of the bus. It also means that
the students that go to a bus stop may not be divided into
groups which may then each take a different bus. A second
assumption is that all buses have equal capacity. Finally, as
mentioned, we assume that each student counts as one unit.
These assumptions may be relaxed in future research.

4. PROBLEM INSTANCE GENERATION

We have designed and implemented an instance genera-

tor for this problem that can generate random problem
instances of any size. The generator requires 5 parame-
ters per instance: np (the number of potential stops), ns
(the number of students per stop), xd , yd (the x and y-
coordinates of the depot) and dmax (the maximum walking
distance).

The instances are generated on the Euclidean square de-

fined by (0, 0) and (xmax , ymax ). It first generates np stops
in this square. The coordinates (xi , yi ) of stop i are uni-
formly distributed in the intervals [wmax , xmax − wmax ] and
[wmax , ymax − wmax ] respectively. In this way, no stu-
dent is ever generated outside the bounderies (0, 0) and
Fig. 3: A medium size problem
The model without any breaking subtour constraints (Con-
straints (4)) is solved and the result is the one obtained on
Figure 4. Clearly, this solution does not satisfy the global

1554
problem constraints and 3 illegal subtours are built (red
dashed arcs on the figure). The value of the objective func-
tion (1) is 245.74. This value is a lower bound of the opti-
mal objective function value.

Fig. 5: Solution after the second iteration

Fig. 4: Solution after the first iteration

Before starting a new iteration, the illegal subtours are au-

tomatically detected and the corresponding contraints are
added to the model. The problem is solved again with these
additional contraints. The results after the second iteration
are shown on Figure 5. Again two illegal subtours are cre-
ated and should be removed during the next iteration.

The objective function value is now 321.23, an improved

lower bound of the optimal objective function value. The
two illegal subtours correspond to two new constraints that
are added to the model and the problem is solved again.
The result of the third iteration is presented on Figure 6.

Now only one illegal subtour remains and the objective

function value is 327.26. After adding the corresponding
contraint to the model, the problem is solved and the opti-
mal solution is presented in Figure 7. The objective func-
tion value is 330.87 and this is the optimal cost for the con-
sidered problem.

Details of the iterations are provided in Table 1. The first

column gives de iteration counter. The second column dis- Fig. 6: Solution after the third iteration
plays the illegal subtours observed at the end of the iter-
ation; (5,9) for example means that an illegal subtour for
stop 5 to stop 9 and back to stop 5 was found. Column It. Subtours Cost CPU (s) ILP it.
“Cost” gives the current value of the objective function. 1 (5,9) (2,10) (4,6) 245.74 1424 129427
The last two columns give the CPU time and the number 2 (2,8) (4,10) 321.23 2966 229393
of Integer Programming iterations of this subtour elimina- 3 (4,7) 327.26 2303 134987
tion iteration. Only four iterations of the subtour elimina- 4 – 330.87 5676 313358
tion procedure are necessary to find the optimal solutions. Optimal solution 330.87 12369 –
But each of these iterations costs a lot since an ILP has to
be optimaly solved. Of course such a methodology can be Tab. 1: Instance 1: details of the iterations
applied only to small or medium size instances.

1555
 It. Subtours Cost CPU (s) ILP it.
1 (6,10) (1,9) (5,8) 294.67 84 9665
2 – 307.44 143 7575
Optimal solution 307.44 227 –

Tab. 3: Instance 3: details of the iterations

teger solutions. For instance 2, the result obtained gives a

total cost of 262.21 which is in fact only 1.98% from the
bound provided at the last complete iteration of Table 2.
15 subtour elimination iterations and 9780 seconds of CPU
time were necessary to reach the end of the search and ob-
tain the first feasible solution. Concerning instance 3, we
were lucky enough to obtain the optimal solution in 227
seconds of CPU time and after only one subtour elimina-
tion iteration. The optimal solution of the subproblem at
the first iteration of the subtour elimination procedure was
found in only 84 seconds of CPU time. Results are reported
in Table 3.

6. CONCLUSIONS AND FUTURE RESEARCH

In this paper, we have proposed an IP formulation for a

schoolbus routing problem in which selection of stops from
Fig. 7: One optimal solution a set of potential stops and allocation of students to stops
were additional decision variables. Our research efforts are
now aimed in two directions: a cutting plane algorithm to
It. Subtours Cost CPU (s) ILP it.
obtain lower bounds of the problem and a metaheuristic
1 (5,8) (2,6) (4,9) 223.90 10686 639683
optimization approach to solve large scale instances. In the
2 (2,9) 235.17 19790 1324610
future, additional features may be added to the formulation,
3 (2,4,9) 236.24 10900 590679
such as allowing more buses to visit a stop, adding time
4 (2,9,4) 236.25 10722 547891
window constraints, etc.
5 (3,9) (2,4) 243.81 6858 482269
6 (6,9) (2,7) 252.57 10406 628056 REFERENCES
7 (3,6) 253.75 7892 426228
8 (4,6) 254.33 9961 563522 [1] B. Bennett and D. Gazis. “School bus routing by com-
9 (3,9,4) 255.90 45045 2173264 puter”. Transportation Research, 6:317–326, 1972.
10 (3,4,9) 255.90 27594 1314457 [2] L. Bodin and L. Berman. “Routing and scheduling of
11 (2,6,4) 257.10 26829 1311801 school buses by computer”. Transportation Science,
12 (2,4,6) 257.10 5712 278011 13:113–129, 1979.
13 Manually stopped after 37369s
Unfinished after about 64 hours of CPU time [3] R. Bowerman, B. Hall, and P. Calamai. “A multiob-
jective optimization approach to urban school bus rout-
Tab. 2: Instance 2: details of the iterations ing: formulation and solution method”. Transportation
Research Part A: Policy and Practice, 29A:107–123,
1995.
Tests have been conducted on other similar instances with
different results. A second instance with the same number [4] J. Braca, J. Bramel, B. Posner, and Simchi-Levi D. “A
of students and stops ran for hours and was stopped before computerized approach to the New York City school
giving a feasible solution. Partial results are reported in bus routing problem”. IIE Transactions, 29:693–702,
Table 2. 1997.
[5] L. Chapleau, J. Ferland, and J. Rousseau. “Cluster-
Approximate solutions can be found by truncating the ing for routing in densely populated areas”. European
search. This is done by setting a stopping condition in Journal of Operational Research, 20:48–57, 1985.
the ILP solver (limit of CPU time between two integer [6] J. Desrosiers, J. Ferland, J.-M. Rousseau, G. Lapalme,
solutions). Since one integer solution corresponds to sev- and L. Chapleau. “An overview of a school busing sys-
eral bus tours (including illegal subtours), we can stop the tem”. In N. Jaiswal, editor, Internal Conference on
search and check for illegal subtours. After adding the cor- Transportation, volume IX of Scientific Management
responding constraints, we can solve the augmented prob- of Transportation, pages 235–243, New Delhi, India,
lem. The cost value in that case has no signification (in 1980.
terms of bounding) but if a feasible solution is obtained at
the end (with no illegal subtours), this is an upper bound [7] G. Dulac, J. Ferland, and P.-A. Fogues. “School bus
of the optimal solution. We have proceed like this for in- routes generator in urban surroundings”. Computers
stances 2 and 3. and Operations Research, 7:199–213, 1980.
[8] S.R. Thangiah, B. Wilson, A. Pitluga, and W. Mennell.
We arbitrarily set the time limit to 600s between two in- “School bus routing in rural school districts”. Working

1556
 paper, Computer Science Department, Slippery Rock
University, Slippery Rock, PA, USA, 2005.
[9] P. Toth and D. Vigo, editors. “The vehicle routing prob-
lem”. SIAM Monographs on Discrete Mathematics
and Applications. Society for Industrial and Applied
Mathematics, Philadelphia, PA, USA, 2001.

1557